thats kind of what i thought the soft start might be
heheh
Soft-Start = Here Be Loads
Posted by renesis at 11:30 | permalink | 0 comments
lordpil's blog
but irl, its going to sink more current and prob slow it down at high slew
yeah but internal load is prob nothing
but its prob fast with nothing
so youre saying its some sort of sample and hold thing? whats it dump through?
more invisible diodes?
maybe it reduces power?
just fill and dump cap instead of constantly pushing through a resistive load
Posted by renesis at 11:25 | permalink | 0 comments
er, does nothing
oh
look, softstart has an arrow to amp's bottom, like supply control, it doesnt output to same node as amp out
test condition is just 1M, so maybe just slows amp down at HF
ideally absolutely nothing
Posted by renesis at 11:20 | permalink | 0 comments
right im saying looks like all the RC is going to do is hold up the signal there for a blip
so basically theres no way youd be abe to fgure this out from 21:17:08 <@BlackMoon> (Cause the comp pin is after all the diodes)
damn
from http://www.ti.com/lit/ds/symlink/lm2585.pdf
page 9 block diagram
no diodes, nothing to indicate pull down op, looks like RC does anything except load amps more at higher freq
Posted by renesis at 11:15 | permalink | 0 comments
blackmoon: so its like, soft off?
looks like all it can do is smooth output turnoff
Posted by renesis at 11:09 | permalink | 0 comments
right so maybe soft start and comp can override the amp output
in jeros other chip it says you can basically shutdown the output by pulling comp pin low
so the comp stuff might just be varying built in soft start
and comp might just be a shitty name
the RC attached to it would start low and keep chip off, no?
wait no its not attachhed like that
shrug
Posted by renesis at 11:03 | permalink | 0 comments
blackmoon: theres a mysterbox at the node just says soft-start
might be simpler than that
just some sort of load that you can tweak
Posted by renesis at 10:58 | permalink | 0 comments
i think they have circuit for that in one of the datasheets
maybe all of them, i thin theyre all basically same datasheet
yeah theres darlington configurations in datasheet, circuit to run a PNP switch that saturates
the darlington circuits burn a lot across the switch
you can maybe do a pfet instead of a pnp but might have to be fast
unless fet is slow and you kill the smps magic
Posted by renesis at 10:06 | permalink | 0 comments
least its cheap
https://octopart.com/search?q=mc34063
cheaper, same part made by TI, ST, and ON
http://www.ti.com/lit/ds/symlink/mc34063a.pdf
no comp pin, has an isense pin
Posted by renesis at 10:00 | permalink | 0 comments
jero32: honestly it doesnt seem to indicate how to load it
without knowing what the soft start circuit is, its hard to predict what it does
searching 'compensa' produces no useful information
they use 1M for test circuits
it might just be the load for the soft start circuit
damn they are not cheap
Posted by renesis at 09:42 | permalink | 0 comments
i want to check out the click stuff from automation direct
cheap, free software
im surprised there isnt more standard MCU boards hardened for industrial automation
Posted by renesis at 05:59 | permalink | 0 comments
kind of yeah, but the associated software is all diff
a lot of them will let you program with C libraries
there are controllers that use scratch programming, interrupt based, but the stuff i used was really broken
throwing up mem allocation errors and shit
like, yo im working with colors and shapes here, i didnt malloc shit, this not my fault
and a non-inuitive menu system
theorectically better than ladder, but implimentation killed it, polish seems a few years out
Posted by renesis at 05:54 | permalink | 0 comments
in general, yes, building in safety factor is reliable, but its also costly
once you know howsomething is going to work, tuning headroom for reliability and cost is a lot of the work
right but models are the same as basic algebra, they get you close
its really hard to model everything typical, and then you have to decide what level of atypical is acceptable
like is it okay if it dies when a vent is covered? if not, how long can it survive with vent covered? should it be aware its going to die and give up?
really depends on the application and liability
most everything compliant either needs to give up or die safe
yup
Posted by renesis at 05:34 | permalink | 0 comments
higher conductivity is better
well, kind of
aluminum, gold are better for moving heat somplace where it han be disspiated better or at least away from sensitive devices
for shielding, something like ceramic doesnt move heat, it takes a long time to heat up
so it actually blocks transfer, like a literal shield, but once it heats up, it takes much longer to cool
do you know how to C/W?
thermal resistance, so temp rise based on power in
right so you want something that moves heat and is configured in a way that the C/W temp rise is acceptable
its not exactly basic algebra, because it depends on orientation of fins for example, and how air can move around
but calculating it gets close enough you can buy something and test, or know you are being reasonably overkill
Posted by renesis at 05:29 | permalink | 0 comments
blackmoon: omg
side move bad, got it
my instinct is usually to push the wood into the guards so guess im doing it as right as possible
Posted by renesis at 02:18 | permalink | 0 comments
Top | Add to Technorati Favorites
© 2007 lordpil. XHTML 1.0! CSS! Site design by GNAA Blog Engine by pbx | MULTI2 | ian hanschen | lolwat